Repository navigation
v2.5.0 — Planning Rigor: Assumptions, Pre-Mortem, Falsification Signals
What's New
New reference file — references/planning-rigor.md adds 7 epistemic techniques to strengthen plans and catch wrong approaches early.
Planning Improvements
| Technique | Where | What it does |
|---|---|---|
| Assumption tracking | plan.md | Every assumption traced to a finding, linked to dependent steps. When one breaks, you know what's invalidated. |
| Pre-mortem & falsification signals | plan.md | "Assume this plan failed — why?" produces concrete STOP IF triggers checked during execution. |
| Exploration confidence | state.md | Quality gate on EXPLORE → PLAN: problem scope, solution space, risk visibility must be "adequate". |
| Prediction accuracy | verification.md | Compare plan predictions vs reality during REFLECT. Calibrates future estimates. |
| Ghost constraint hunting | RE-PLAN | Before pivoting, check if the constraint behind the failed approach is still valid. |
| Decomposition analysis | iteration 5 | Identify 2-3 independent sub-goals before the iteration 6 hard stop. |
| Phase balance heuristic | all states | Rough effort budgets with warning signs for imbalance. |
Other Changes
- "Risks" section removed from plan.md — subsumed by Failure Modes (dependencies) + Pre-Mortem (approach validity).
- Step annotations —
[RISK: low/medium/high]and[deps: N,M]recommended on each plan step.
Design Decisions
- Pre-Mortem and Falsification Signals merged into one section (they produce the same insight).
- Assumptions use a lightweight bullet list, not a table (less boilerplate over 100+ runs).
- Exploration Confidence is a transition log entry, not a file section (ephemeral meta-assessment, not a codebase finding).
See CHANGELOG.md for full details.